Failure of Rituximab in Immune Thrombocytopenia Is Associated with the Activation of Splenic CD8 T Cells

Abstract 623 Background: Different mechanisms are involved in the pathogenesis of immune thrombocytopenia (ITP). Platelet production is impaired as result of an immune response directed against megakaryocytes and an inappropriate level of thrombopoietin. Furthermore, platelets targeted by autoantibodies against their membrane glycoproteins are cleared by the reticuloendothelial system, mainly in the spleen. Thus, treatments targeting the autoimmune humoral response, notably rituximab (off-label use), have been proposed. However, the one-year response rate after rituximab (RTX) is not greater than 40% and the mechanisms involved in its failure remain to be defined. CD8 T cells which have been demonstrated to be involved in platelet destruction both in a mouse ITP model and in humans may play a part in the inefficiency of RTX. Purpose: To compare the splenic CD8 T cell response of ITP patients who have been refractory to RTX with untreated ITP patients Methods: The spleens of 23 primary ITP patients were assessed: 12 patients previously treated with RTX without improvement were examined as RTX refractory patients and 11 patients who have not received RTX were referred as untreated patients. Splenectomy was performed at a median of 7.1 months after RTX administration. Nine post-traumatic spleens were used as controls. Flow cytometry analysis was performed on CD8+ splenocytes to determine the expression of activation marker (HLA-DR), memory T cell markers (CD27, CD28), chemokine receptor (CCR7), adhesion molecule (CD62L) and intracellular cytotoxic proteins (granzyme B, perforin). Lineage commitment of T cells was determined after stimulation for 4 hours with PMA and ionomycin in presence of brefeldin-A, by intracellular staining of interferon-γ (Tc1, Th1), IL-4 (Tc2, Th2) and IL-17 (Tc17, Th17). Results are expressed by median and [interquartile range]. Statistical analysis was performed using non parametric tests (Kruskal-Wallis and Mann-Whitney) to compare the different groups. Results: After RTX infusion, B cells represented only 1.4% [0.3–3.4] in total splenic lymphocytes compared to about 35% in controls and untreated patients. No alteration in the CD8/CD4 ratio was observed after RTX. Nevertheless, expression of HLA-DR and granzyme B by splenic CD8 T cells was increased in RTX refractory patients when compared to untreated patients: 60% [50.6–68.9] vs. 41.8% [33.7–54.2] (p=0.02) and 50.3% [32.2–72.3] vs. 23.9% [14.4–27.7] (p=0.006), respectively. Upon stimulation, interferon-γ expression was significantly higher in CD3+CD8+ (68% [59.1–81.6]) but also in CD3+CD8− (32.5% [24.8–44]) in RTX refractory patients compared to untreated patients. The percentage of CCR7−CD62L− and CD27−CD28− among CD8+ T cells was increased in RTX refractory patients compared to untreated patients, respectively 85.4%[77.1–93.2] vs. 66.5%[56.2–73.2] (p=0.001) and 33.2% [23.3–67.1] vs. 13.4%[11.6–24.9] (p=0.04). Conclusions: Our results show that splenic CD8 T cells from RTX refractory patients express more HLA-DR, granzyme B and interferon-γ, whereas CCR7, CD62L, CD27 and CD28 are lower in comparison to untreated patients. This phenotype is consistent with effector T cells localizing in the red pulp as they lack CCR7 expression, and presumably playing a main role in splenic platelet destruction. Our results strongly support that platelet destruction is preferentially mediated by CD8 T cells rather than by the humoral response in some ITP patients, which may explain their unresponsiveness to rituximab. Disclosures: Off Label Use: Rituximab used during Immune Thrombocytopenia.